How to read your AI visibility

0:00Okay, now let's move to the second part, which is actually understanding your visibility scores. Learning how you read your scores is very important for you to take on the next step, which is creating the work for it. The next step is actually the analysis, reading your numbers, understanding the scores. We've set up the prompt in the first part. We've let the prompts run, and now we want to look at the results of these prompts. When the AI are asked these prompts, what are the results that came up to them?

0:28Before I show you everything on Hopkins dashboard, there's few very important words that you need to understand. Words that are very important for you to learn. First is the visibility score. Visibility score is very easy to understand. It's a number that tells you how often you show up inside of AI answers. So let's say one answer here is being asked to the AI model.

0:56What are the best shoes to buy? And your brand is setting shoes. Let's say when this question is being asked 10 times to an AI model, you're only showing up 2 times. That simply means your visibility score or your visibility rate is around 20%. Now, what is the second part, which is the share of voice? This is your slice of mentions compared to the competitors.

1:28Across those two times that you showed up, you're 20%. Your competitors probably showed up way more than you. The difference between that is your share of voice percentage. Now, the third part is actually very interesting, which is your ranking inside of these answers. Just say you're being named. You're being named 20% of the times.

1:53But where exactly inside of the answer are you getting named? Usually in commercial questions, the AI tries to rank the options of every brand that's naming. We track where exactly are you being named and where exactly are being ranked inside of these answers. And the last thing is the citation. Did this answer pull one of your page links or not? Now, I also want you to understand the difference between the mention versus the citation, where your brand, yes, it could be mentioned, but that does not necessarily mean that it is being cited or your brand could be cited, but that does not necessarily mean that it's being mentioned.

2:41If your brand is being mentioned and also being cited at the same time, that means that the AI looked for the sources online, picked your page, found the answer to the question that is trying to answer, and it chose to pull your brand as part of this answer and mention you inside the question output that the person was asking about. But what happens if your brand is being cited, so it is using the links of your pages, but it's not mentioning you? That probably means that the AI is finding something interesting in your page, but it still is not enough for it to be mentioned. Let's look at all of this inside of Hopkins dashboard.

3:31If you come here, this is the main view of our dashboard. Underneath the visibility score inside the overview, this is where you're able to track all of your most important metrics, which are the visibility scores, the chair of voice, and your average position. You're able also to attract the visibility scores compared to the competitors that you see here and where you fall in terms of your visibility ranking. You can even change this view across the different model, and this is the most interesting part that helps you identify the gaps. If you look here at the engines, you have different tools of AI that you can search between. If you use ChatGPT, for example, to find out your visibility score inside of ChatGPT, you'll find that it could be different from your visibility score inside of Claude, and the same goes with each and every single model.

4:27Every model sees you differently. You can show up in a single model, but not show up in another. This is what this dashboard is used for. If you use ChatGPT, for example, you can even take this to the next level to understand across ChatGPT your visibility scores, your chair of voice, and your average position in each and every single answer. Here as well, you can look at different topics, and this is where it gets interesting. This is how you identify the gaps.

4:59You have different topics that I helped you set up previously. Now you want to look at each and every single topic of them. Are you showing up in topics, and are you showing up in some topics and some other topics you're not showing up? Maybe you show up in a different product category topic that you have, but you're not showing up in another product category topic that is out there. This is how you help identify this. So let's say if we pick this topic, for example, here our visibility score is around 4.

5:315%, which is relatively low. But if you look at, for example, another topic, which is AI visibility tools, you'll find that this topic has also only 0% visibility. So now you're able to easily identify that this specific topic, your visibility in it is extremely low, and this is the first gap that we're able to find from here. Now let's take this into a much deeper look. If you move to the citations, and the citations is what helps you identify where the AI looked for to find its sources. So the AI went online, it looked at multiple different sources, and these are the pages, the domains, that it chose to cite its information from to give an answer to the user.

6:15If you take a look here, you'll find that inside of chatGPT across this specific topic, the most dominant part of the citation was actually TechCradar, which is a blog post. You'll find other different sources from which it pulled its information from. But you can see here that there's a very clear direction between the AI looking for sources through the competitors, and as well as through the earned media. So what does this mean? This means that you need to build a very strong onsite pages that are answering the exact questions that are being asked inside the topic of AI visibility tool, but that would not just be enough. You would also need to show up inside of TechCradar, you need to show up inside of Reddit, and other subreddits.

7:12If you go down there, you'll take a look at the most cited domains here. These are the domains that AI trusted the most to pull its information from. This bar here, you'll find that these most cited domains probably are a mix of competitors and earned media, and this is what you need to focus on. If you take a look as well at the top cited pages, you will be able to identify the exact pages that the AI pulled from its answers. So for you to use these pages, to later use them to identify what exactly is the content that you need to be shopping. Why?

7:51Because this is the exact content that the AI is looking for. But don't forget, your content still needs to be insightful, and it still needs to be unique only to you. But not also that. You would also need to be published inside of the off-site media, which is, let's say for example, TechCradar, Reddit, and other different sources. Now, there's something very important that a lot of people miss on, which is while you're interpreting your questions, you need to make sure that these questions are actually being read correctly. And you do that through the settings page, where here you're able to track your right competitors.

8:29So here you have a list of competitors. You need to make sure that the aliases are correct of these competitors, and that you're tracking the correct competitors. You can do that through here very easily on the name matching, where you will find a list of many different competitors, and you choose whether the competitor that is being tracked here is a correct competitor, or is it not a correct competitor. You can also choose to update your citation categories, which is the list of the domains that the AI chooses to cite from, to make sure that you're able to read the citation answers much more accurately. Don't forget this part. This part is very important because it helps you actually identify and rank better, and shows you the answers in a more correct format.

9:14Now, for the next part of the video, and our next video, I will show you what you can do after actually finding out the gap and the problems that you just saw here.